The Size of Settlement and Urbanization in Sindh

Rural to urban migration is one route to urbanization. Another one is the increase in size and change in
complexity of smaller rural settlements, as they acquire urban characteristics. This has been ably
documented by Raza Ali in the Pakistani context particularly with respect to Punjab. This note will show
that the size and pattern of settlements in rural Sindh is considerably different, with relatively small
settlements accounting for a large proportion of settlements and the population. The tendency towards
‘fragmented’ settlements, therefore, needs to be juxtaposed alongside the more conventional trend
towards consolidation. There are distinct historical and sociological explanations for this settlement
pattern, as well as clear implications for the trajectories of urbanization in Sindh. The note will illustrate
this issue as well as the challenges for reliable documentation, data collection, and programming, using
case studies of individual revenue villages in Sindh. Implications for research as well as policy-making
will be spelled out.

What do we know about rural settlements in Pakistan? And what do we know about them in Sindh? The
most common way of knowing about settlements is the population census. The census enumerates the
population using territorial divisions which are based on administrative boundaries. One important
division is between urban and rural areas. Urban areas are usually identified as those territories which
fall under municipal governance of various types. For rural areas the census relies on the territorial
demarcation of the land revenue department, which, historically, formed the basis for civil
administration over much of the country.

In Sindh the deh is the basic administrative and revenue unit, and is the lowest territorial unit for which
the population census publishes data. In other provinces this basic rural unit is called the mouza. The
deh or mouza is the used to identify the lowest population agglomeration for virtually every social policy
and research purpose. Moreover, the deh or mouza sometimes end up being used as synonyms for
rural settlements. This works, to some extent, in many parts of Punjab and some regions of Khyber
Pakhtunkhwa where the rural settlement often bears the name of the mouza, and much of the
population of the mouza resides in a geographically compact agglomeration.

But this is not the case in Sindh. There is little correspondence here between the identity of the deh and
that of the goth which is how communities refer to their villages or settlements. A deh can have over a
dozen goths of various sizes spread over its territory. The goth is usually made up of smaller segments
known as paro (plural para) which are generally inhabited by close kin. Some of the smaller goths are
just single paro settlements, while the bigger ones can have several para. In any case, the deh does not
tell us much about the actual rural settlement. The Sindh government has conducted goth surveys from
time to time. One such survey conducted in 1990 found that there were 3,643 goths in the 452 dehs of
Dadu district, 1,765 goths in the 281 dehs of what was then Nawabshah (now Shaheed Benazirabad),
and 2,863 goths in Hyderabad’s 374 dehs. As we show later, goths enumerated in the survey are those
which were thought to be eligible for government recognition – numerous smaller goths are not
included in the list.

In a study of rural settlements in different regions of Pakistan we carried out settlement-wise population
censuses in seven administrative villages (deh in Sindh and mouza elsewhere) in 2005 (Table 1). There
was great variation between these administrative villages in the geographic spread of the population
within their boundaries. In the fieldwork site in Toba Tek Singh district in central Punjab which had a
population of over 2,500, the entire population resided in one consolidated settlement. In Thatta in
southern Sindh, however, our fieldwork deh of 1,829 persons consisted of 22 distinct settlements of

1
which 16 had fewer than 100 people. The settlement pattern was found to be similarly fragmented in
Muzzaffargarh in southern Punjab and in Upper Dir in northern Khyber Pakhtunkhwa.

Table 1: Settlement patterns in 7 administrative villages across Pakistan

District of Population No of Average No of No of
administrative settlements per settlements settlements
village where settlement with > 400 with < 100
fieldwork people people
conducted
Chakwal 491 2 246 1 1
Toba Tek Singh 2,638 1 2,638 1 0
Muzzaffargarh 1,878 11 171 1 5
Upper Dir 1,171 13 90 0 9
Mardan 1,164 3 388 2 1
Sanghar 2,999 12 250 1 4
Thatta 1,829 22 83 0 16
Source: Gazdar, Haris (2006), ‘Rural Economy and Livelihoods’, Asian Development Bank, Islamabad.

Turning back to Sindh, the goth is not only distinct from the deh which is the land revenue system’s
administrative unit. It is also a very different type of settlement from the traditional raj which was a
relatively large composite settlement made of multiple classes, kinship groups and occupations. The
Sindhi raj, in fact, was similar to the historical pattern of the village republic found in Punjab and other
parts of South Asia. The raj combined an economic system with a closely linked governance structure -
with landlords and farmers at the top, the raj had artisans, craftsmen, and other non-agricultural
occupations, all woven together into one whole. Many of the older raj settlements have already grown
into small towns and lost their rural character. But unlike Punjab, Sindh has also seen a pattern of
settlement fragmentation towards goths.

The question arises how do we know what the rural settlement in Sindh looks like. How many are there,
and what are their sizes in terms of population? There are at least four distinct ways in which the
government system attempts to know these rural settlements. First, at the local level there is the land
revenue cadaster known as the Mukhtiarkar map. Settlements are marked on this as areas which are
formally designated as inhabited zones from the land revenue point of view. This does not necessarily
mean that all actual settlements are covered, or that all areas formally marked as inhabited zones
actually have settlements on them. Second, there is the Sindh Goth Abad Authority’s village directory
which is based on a survey carried out in 1990 for the implementation of the Sindh Goth Abad act.i This
list only includes settlements thought to have at least 50 households. Third is a Rural Settlements
Survey carried out by the Sindh Bureau of Statistics – sporadically in particularly districts – which
includes on those settlements that are thought to have more than 200 households. Finally there is the
Sindh Education Management Information System or SEMIS which maintains a record of all government
schools in the province. The school record includes information on the UC and deh where the school is
located.

2
 Table 2: Settlements in Deh Bhaji, District Sanghar
Collective census SGAS SBS
Settlement Population Households Estimated Estimated
population households
1 Bhaji 1,080 145 300 300
2 M A Bagrani 393 55 200 80
3 W Bagrani 349 41 250
4 M Rind 234 31 150
5 G Bagrani 202 28 200 30
6 Jaityo Bheel 189 27
7 B Wassan 148 19 150
8 S Kerio 144 16 300
9 D Bheel 108 11
10 Khaskheli 82 10 200
11 Sanjrani 70 8 150 300
12 Maachhi 0 0
Total 2,999 391 1,900 710
Source: Collective fieldwork 2005; Sindh Goth Abad Survey; Sindh Bureau of Statistics

For one deh in the Sanghar district of central Sindh we were able to compare the settlement record from
all of these sources. In addition, here we were able to carry out a complete house-to-house population
census ourselves and identify each distinct settlement. A comparison between our own census, the
SGAS list and the SBS findings is presented in Table 2. While we had found 11 distinct settlements and
one abandoned settlement (Maachhi) the SGAS directory had missed two: both belonging the non-
Muslim and socially marginalized Bheel community. The SGAS had also missed the Maachhi settlement
although it had been inhabited in 1990 and was abandoned only in 1997. The SBS list was even more
restricted. It had only 4 settlements. The omissions in both these lists were not necessarily driven by
their own cut-off rules. The Sanjrani settlement was in the SBS list although it had only 8 households.
One of the largest landlords of the deh happens to reside in this village. Similarly the SGAS list omitted
the two Bheel villages while in included some villages that were even smaller than these villages. There
were also other discrepancies. The SGAS list included another five goths with an additional population
of 900 persons which were not traceable on the ground. There were also discrepancies between the
SEMIS list and our own census, as well as other sources. SEMIS lists a school in the Sanjrani village while
none exists on the ground. It also lists a school in Sultan Bhanoojo which happens to be in a
neighbouring deh as belonging to Deh Bhaji.

The fragmented settlement phenomenon in rural Sindh has implications not only for accurate
measurement. Important social policy issues are at stake. For a population of 10 million children aged
5-16 Sindh had over 42,000 government schools, compared with Punjab which served 23 million
children through 57,000 schools. The difference is almost entirely due to the smaller settlement size in
rural Sindh compared with rural Punjab. The Sindh schools had an average of 3.5 teachers compared
with 5.8 in Punjab.ii

There are implications of Sindh’s distinctive pattern of rural settlement for the process of urbanization.
Part of the historical trend of village consolidation into towns which is observed in Punjab (e.g. in Raza
Ali’s work) is visible in Sindh with respect the conversion of the old raj villages into towns. There is also,

3
however, a strong countercurrent in the shape of fragmentation of settlements along kinship group lines
into smaller goths and para. This countercurrent is driven by the desire and ability of kinship groups to
acquire relative autonomy from their more powerful neighbours. This can be observed even among
historically marginalized groups who, by establishing their villages, acquire a stronger sense of identity
and political power. There is an active contest for the recognition of a settlement and for the acquisition
of public infrastructure such as schools and electricity. These public goods are valued not only in
themselves, but also as a way of creating greater stability and security of tenure for the settlement. This
trajectory is remarkably similar to the process of regularization of unplanned urban settlements in urban
Sindh.iii

An understanding of the rural settlement is important for making sense of urbanization in Sindh. The
fragmented settlement not only poses identification, measurement and social policy challenges, it also
influences the pace of urbanization by slowing down the agglomeration of rural settlements into towns.
At the same time, however, there appear to be many points of similarity in the political economy of
rural settlement fragmentation on the one hand and the consolidation of urban unplanned settlements
on the other. Going forward, there is an urgent need for the more precise identification and
documentation of rural settlements, using both ethnographic methods but also modern techniques in
spatial mapping. While we have provided observations based on limited primary fieldwork and
comparison with government records and lists, Sindh needs to undertake a far more extensive exercise
in the accurate mapping of settlements and their populations. This is crucial not only for social policy
and service delivery, but for a more grounded analysis of urbanization in Sindh.

i
For details about the Sindh Goth Abad scheme and its implementation see Gazdar, Haris and Hussain Bux Mallah,
“Residential Security as Social Protection”, in IDS Bulletin, July 2010.
ii
These figures are based on authors’ calculations from the Pakistan Standard of Living Measurement Survey.
iii
We have argued this elsewhere, see for example, Gazdar, Haris and Hussain Bux Mallah, “The Making of a
'Colony' in Karachi and the Politics of Regularisation”, South Asia Multidisciplinary Academic Journal, Thematic
Issue No. 5, Rethinking Urban Democracy, February 2012.